A method for edge online processing of multi-source data

Abstract

An edge online processing method for multi-source data relates to the technical field of data processing, and obtains data corresponding to various types of production data monitoring indicators of process subsequences; constructs a relationship topology diagram between industrial equipment corresponding to each process subsequence, obtains abnormal process subsequences according to an input edge data fuzzy judgment model, suspends the operation of production equipment of the abnormal process subsequence, and starts an edge scheduling mechanism for the abnormal process subsequence; constructs a full-process edge data prediction model based on a graph convolutional neural network, obtains prediction data corresponding to various types of production data monitoring indicators of other edge monitoring nodes according to the full-process edge data prediction model, and judges whether other edge monitoring nodes are abnormal process subsequences to be detected, thereby realizing the prediction of early faults before industrial equipment changes from a normal state to a fault state, thereby taking preventive measures before the industrial equipment fault matures.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an edge online processing method for multi-source data. Background Art

[0002] In the existing industrial production equipment fault warning technology, threshold monitoring technology is often used for various indicators of production equipment (equipment operating status, temperature, vibration). This is a static analysis method. Early faults before industrial equipment changes from a normal state to a faulty state are often difficult to detect, and the advantages of online monitoring systems are not reflected. Because the signs of industrial equipment failure may be weak or difficult to identify, this may lead to the inability to take preventive measures before the failure matures. In order to solve the above technical problems, an edge online processing method for multi-source data is provided. Summary of the invention

[0003] In order to solve the above technical problems, the object of the present invention is to provide an edge online processing method for multi-source data, comprising the following steps:

[0004] Step S1: According to the production plan information, the production process is divided into several process sub-sequences, edge monitoring points are set in each process sub-sequence, and data corresponding to each type of production data monitoring indicator of the process sub-sequence is obtained;

[0005] Step S2: construct a relationship topology diagram between the industrial equipment corresponding to each process subsequence, set an edge data fuzzy judgment model at the edge monitoring point of each process subsequence, obtain the abnormal process subsequence according to the input edge data fuzzy judgment model, suspend the operation of the production equipment of the abnormal process subsequence, and start the edge scheduling mechanism of the abnormal process subsequence;

[0006] Step S3: Build a full-process edge data prediction model based on the graph convolutional neural network, obtain the prediction data corresponding to each type of production data monitoring indicator of other edge monitoring nodes according to the full-process edge data prediction model, and determine whether each other edge monitoring node is an abnormal process subsequence to be detected.

[0007] Furthermore, according to the production plan information, the production process is divided into several process sub-sequences, edge monitoring points are set in each process sub-sequence, and the process of obtaining data corresponding to each type of production data monitoring indicator of the process sub-sequence includes:

[0008] Acquire the production plan information of the industrial production process of the current collection cycle, acquire the working time period and working order of each industrial equipment according to the production plan information, construct the production process sequence of the current collection cycle according to the working time period and working order of each industrial equipment, and segment the production process sequence according to the working time period of different industrial equipment in the production process sequence, and divide it into several process subsequences;

[0009] Set edge monitoring points in each process subsequence to obtain the process characteristics of each industrial equipment, and use data retrieval to obtain various types of production data monitoring indicators of each edge monitoring point according to the process characteristics of each industrial equipment;

[0010] The edge monitoring points acquire data corresponding to various types of production data monitoring indicators in real time according to various types of production data monitoring indicators, mark the collection time, and set the collection cycle.

[0011] Furthermore, the process of constructing a relationship topology diagram between industrial equipment corresponding to each process subsequence includes:

[0012] Obtain the process flow characteristics of each industrial equipment, obtain the dependency relationship between the industrial equipment of each process subsequence according to the process flow characteristics of each industrial equipment, take each process subsequence as a node, take the dependency relationship between the industrial equipment of each process subsequence as the connection relationship between the nodes, and construct a relationship topology diagram;

[0013] A local database is set at the edge monitoring point of each process subsequence, and the relationship topology map is stored in the local database of the data monitoring point of each process subsequence.

[0014] Furthermore, an edge data fuzzy judgment model is set at the edge monitoring point of each process subsequence, and the process of obtaining an abnormal process subsequence according to the input edge data fuzzy judgment model includes:

[0015] An edge data fuzzy judgment model is set at the edge monitoring point of each process subsequence, and the data threshold interval corresponding to each type of production data monitoring indicator of each process subsequence is obtained in advance. The data threshold interval corresponding to each type of production data monitoring indicator of each process subsequence is taken as the fuzzy adaptive rule of the edge data fuzzy judgment model, and the data corresponding to each type of production data monitoring indicator collected in real time by the edge monitoring point of the process subsequence at the current moment is obtained. The data corresponding to each type of production data monitoring indicator is input into the edge data fuzzy judgment model of the edge monitoring point, and the edge data fuzzy judgment model compares the data corresponding to each type of production data monitoring indicator with the data threshold interval corresponding to each type of production data monitoring indicator;

[0016] If the data corresponding to each type of production data monitoring indicator falls within the corresponding data threshold range, it proves that the production process of the industrial equipment corresponding to the process subsequence is normal;

[0017] If the data corresponding to a certain type of production data monitoring indicator does not fall within the corresponding data threshold interval or the data corresponding to multiple types of production data monitoring indicators do not fall within the corresponding data threshold interval, the process subsequence is marked as an abnormal process subsequence.

[0018] Furthermore, the process of suspending the operation of the production equipment of the abnormal process subsequence and starting the edge scheduling mechanism of the abnormal process subsequence includes:

[0019] The monitoring center is communicated with the edge monitoring nodes of each process subsequence. When a process subsequence is marked as an abnormal process subsequence, the edge monitoring nodes of the abnormal process subsequence send abnormal monitoring indicator data to the monitoring center, and the monitoring center suspends the operation of the production equipment of the abnormal process subsequence. Subsequently, the edge monitoring points of the abnormal process subsequence obtain other process subsequences that have no connection relationship with the abnormal process subsequence according to the relationship topology diagram in the local database, obtain the remaining working period duration of the abnormal process subsequence after the suspension of operation and the working period duration of other process subsequences, screen out other process subsequences whose working period duration is less than or equal to the remaining working period duration of the abnormal process subsequence, randomly select one other process subsequence from the other process subsequences and mark it as the first process subsequence, obtain the working period and working order of the first process subsequence, and exchange the working period and working order of the abnormal process subsequence with the working period and working order of the first process subsequence.

[0020] Furthermore, a full-process edge data prediction model is constructed based on a graph convolutional neural network. The process of obtaining prediction data corresponding to various types of production data monitoring indicators of other edge monitoring nodes according to the full-process edge data prediction model includes:

[0021] The monitoring center stores historical data corresponding to various types of production data monitoring indicators stored in several historical collection cycles of each edge monitoring node, and uses the historical data corresponding to various types of production data monitoring indicators stored in several historical collection cycles of each edge monitoring node as a supplementary node to the node to which each edge monitoring node belongs in the relationship topology diagram;

[0022] The influence degree of dependency relationship on other nodes is calculated according to the adjacency matrix of each node in the relationship topology graph, and the weight label is set for the node by obtaining the influence degree of dependency relationship of each node on other nodes;

[0023] Subsequently, a full-process edge data prediction model is constructed based on the graph convolutional neural network, the relationship topology graph and the supplementary nodes in the relationship topology graph are learned, and the data corresponding to each type of production data monitoring indicators of the abnormal process subsequence are input into the learned full-process edge data prediction model. According to the full-process edge data prediction model, the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes are obtained, and the monitoring center sends the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes to the corresponding edge monitoring nodes.

[0024] Further, according to the prediction data corresponding to each type of production data monitoring indicator of each other edge monitoring node, the process of judging whether each other edge monitoring node is an abnormal process subsequence to be detected includes:

[0025] Each other edge monitoring node inputs the received prediction data corresponding to each type of production data monitoring indicator into the edge data fuzzy judgment model contained in each other edge monitoring node, and the edge data fuzzy judgment model compares the prediction data corresponding to each type of production data monitoring indicator with the data threshold interval corresponding to each type of production data monitoring indicator;

[0026] If the predicted data corresponding to a certain type of production data monitoring indicator does not fall within the corresponding data threshold interval or the predicted data corresponding to multiple types of production data monitoring indicators do not fall within the corresponding data threshold interval, the process subsequence is marked as an abnormal process subsequence to be detected.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. Construct a relationship topology diagram between the industrial equipment corresponding to each process subsequence, set up an edge data fuzzy judgment model at the edge monitoring points of each process subsequence, obtain the abnormal process subsequence according to the input edge data fuzzy judgment model, suspend the operation of the production equipment of the abnormal process subsequence, and start the edge scheduling mechanism of the abnormal process subsequence to adapt to actual production needs and changes, reduce the impact of equipment failures on the industrial production process, and improve industrial production efficiency.

[0029] 2. Build a full-process edge data prediction model based on the graph convolutional neural network, and obtain the prediction data corresponding to each type of production data monitoring indicator of each other edge monitoring node according to the full-process edge data prediction model. According to the prediction data corresponding to each type of production data monitoring indicator of each other edge monitoring node, judge whether each other edge monitoring node is an abnormal process subsequence to be detected. When the process subsequence is marked as an abnormal process subsequence to be detected, the edge monitoring node of the abnormal process subsequence to be detected sends an abnormal detection instruction to the monitoring center, and the monitoring center arranges personnel to perform early maintenance on the industrial equipment corresponding to the abnormal process subsequence to be detected, thereby realizing the prediction of early faults before the industrial equipment changes from a normal state to a fault state, so as to take preventive measures before the industrial equipment fault matures. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a schematic diagram of a method for online edge processing of multi-source data according to an embodiment of the present application. DETAILED DESCRIPTION

[0031] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0032] like Figure 1 As shown, a method for edge online processing of multi-source data includes the following steps:

[0033] Step S1: According to the production plan information, the production process is divided into several process sub-sequences, edge monitoring points are set in each process sub-sequence, and data corresponding to each type of production data monitoring indicator of the process sub-sequence is obtained;

[0034] Step S2: construct a relationship topology diagram between the industrial equipment corresponding to each process subsequence, set an edge data fuzzy judgment model at the edge monitoring point of each process subsequence, obtain the abnormal process subsequence according to the input edge data fuzzy judgment model, suspend the operation of the production equipment of the abnormal process subsequence, and start the edge scheduling mechanism of the abnormal process subsequence;

[0035] Step S3: Build a full-process edge data prediction model based on the graph convolutional neural network, obtain the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes according to the full-process edge data prediction model, and judge whether each other edge monitoring node is an abnormal process subsequence to be detected based on the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes.

[0036] It should be further explained that, in the specific implementation process, according to the production plan information, the production process is divided into several process sub-sequences, edge monitoring points are set in each process sub-sequence, and the process of obtaining data corresponding to each type of production data monitoring indicator of the process sub-sequence includes:

[0037] Acquire the production plan information of the industrial production process of the current collection cycle, acquire the working time period and working order of each industrial equipment according to the production plan information, construct the production process sequence of the current collection cycle according to the working time period and working order of each industrial equipment, and segment the production process sequence according to the working time period of different industrial equipment in the production process sequence, and divide it into several process subsequences;

[0038] Set edge monitoring points in each process subsequence to obtain the process characteristics of each industrial equipment, and use data retrieval to obtain various types of production data monitoring indicators of each edge monitoring point according to the process characteristics of each industrial equipment;

[0039] The edge monitoring points acquire data corresponding to various types of production data monitoring indicators in real time according to various types of production data monitoring indicators, mark the collection time, and set the collection cycle.

[0040] It should be further explained that, in the specific implementation process, the process of constructing the relationship topology diagram between the industrial equipment corresponding to each process subsequence includes:

[0041] Obtain the process flow characteristics of each industrial equipment, obtain the dependency relationship between the industrial equipment of each process subsequence according to the process flow characteristics of each industrial equipment, take each process subsequence as a node, take the dependency relationship between the industrial equipment of each process subsequence as the connection relationship between the nodes, and construct a relationship topology diagram;

[0042] The influence degree of dependency relationship on other nodes is calculated according to the adjacency matrix of each node in the relationship topology graph, and the weight label is set for the node by obtaining the influence degree of dependency relationship of each node on other nodes;

[0043] It should be further explained that the dependency relationship between the industrial equipment of each process subsequence means that the working time or performance of the industrial equipment of the process subsequence may affect the working efficiency of other mining equipment;

[0044] A local database is set at the edge monitoring point of each process subsequence, and the relationship topology map is stored in the local database of the data monitoring point of each process subsequence.

[0045] It should be further explained that, in the specific implementation process, an edge data fuzzy judgment model is set at the edge monitoring point of each process subsequence, and the process of obtaining an abnormal process subsequence according to the input edge data fuzzy judgment model includes:

[0046] An edge data fuzzy judgment model is set at the edge monitoring point of each process subsequence, and the data threshold interval corresponding to each type of production data monitoring indicator of each process subsequence is obtained in advance. The data threshold interval corresponding to each type of production data monitoring indicator of each process subsequence is taken as the fuzzy adaptive rule of the edge data fuzzy judgment model, and the data corresponding to each type of production data monitoring indicator collected in real time by the edge monitoring point of the process subsequence at the current moment is obtained. The data corresponding to each type of production data monitoring indicator is input into the edge data fuzzy judgment model of the edge monitoring point, and the edge data fuzzy judgment model compares the data corresponding to each type of production data monitoring indicator with the data threshold interval corresponding to each type of production data monitoring indicator;

[0047] If the data corresponding to each type of production data monitoring indicator falls within the corresponding data threshold range, it proves that the production process of the industrial equipment corresponding to the process subsequence is normal;

[0048] If the data corresponding to a certain type of production data monitoring indicator does not fall within the corresponding data threshold interval or the data corresponding to multiple types of production data monitoring indicators do not fall within the corresponding data threshold interval, the process subsequence is marked as an abnormal process subsequence.

[0049] It should be further explained that, in the specific implementation process, the process of suspending the operation of the production equipment of the abnormal process subsequence and starting the edge scheduling mechanism of the abnormal process subsequence includes:

[0050] The monitoring center is communicated with the edge monitoring nodes of each process subsequence. When a process subsequence is marked as an abnormal process subsequence, the edge monitoring nodes of the abnormal process subsequence send abnormal monitoring indicator data to the monitoring center. The monitoring center suspends the operation of the production equipment of the abnormal process subsequence and arranges personnel to repair the equipment of the abnormal process subsequence. Subsequently, the edge monitoring points of the abnormal process subsequence obtain other process subsequences that have no connection relationship with the abnormal process subsequence according to the relationship topology diagram in the local database, obtain the remaining working period duration of the abnormal process subsequence after the suspension of operation and the working period duration of other process subsequences, screen out other process subsequences whose working period duration is less than or equal to the remaining working period duration of the abnormal process subsequence, randomly select one other process subsequence from the other process subsequences and mark it as the first process subsequence, obtain the working period and working order of the first process subsequence, and exchange the working period and working order of the abnormal process subsequence with the working period and working order of the first process subsequence.

[0051] It should be further explained that, in the specific implementation process, a full-process edge data prediction model is constructed based on a graph convolutional neural network, and the process of obtaining prediction data corresponding to various types of production data monitoring indicators of other edge monitoring nodes according to the full-process edge data prediction model includes:

[0052] The monitoring center stores historical data corresponding to various types of production data monitoring indicators stored in several historical collection cycles of each edge monitoring node, and uses the historical data corresponding to various types of production data monitoring indicators stored in several historical collection cycles of each edge monitoring node as a supplementary node to the node to which each edge monitoring node belongs in the relationship topology diagram;

[0053] Subsequently, a full-process edge data prediction model is constructed based on the graph convolutional neural network, the relationship topology graph and the supplementary nodes in the relationship topology graph are learned, and the data corresponding to each type of production data monitoring indicators of the abnormal process subsequence are input into the learned full-process edge data prediction model. According to the full-process edge data prediction model, the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes are obtained, and the monitoring center sends the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes to the corresponding edge monitoring nodes.

[0054] It should be further explained that the calculation formula for calculating the degree of influence of the dependency relationship on other nodes based on the adjacency matrix of each node in the relationship topology graph is:

[0055] ;

[0056] in, , Respectively represent the influence degree of the dependency relationship of a process subsequence on other process subsequences after the i+1th and ith iterations; Y= , whose specific value is obtained through the node adjacency matrix; represents the damping factor; represents the state transfer matrix; E represents an N×1 matrix whose elements are all 1; N represents the number of process subsequences.

[0057] It should be further explained that, in the specific implementation process, according to the prediction data corresponding to the various types of production data monitoring indicators of the other edge monitoring nodes, the process of judging whether the other edge monitoring nodes are abnormal process subsequences to be detected includes:

[0058] Each other edge monitoring node inputs the received prediction data corresponding to each type of production data monitoring indicator into the edge data fuzzy judgment model contained in each other edge monitoring node, and the edge data fuzzy judgment model compares the prediction data corresponding to each type of production data monitoring indicator with the data threshold interval corresponding to each type of production data monitoring indicator;

[0059] If there is a predicted data corresponding to a certain type of production data monitoring indicator that does not fall within the corresponding data threshold interval or the predicted data corresponding to multiple types of production data monitoring indicators do not fall within the corresponding data threshold interval, the process subsequence will be marked as an abnormal process subsequence to be detected. When the process subsequence is marked as an abnormal process subsequence to be detected, the edge monitoring node of the abnormal process subsequence to be detected sends an abnormal detection instruction to the monitoring center, and the monitoring center arranges personnel to perform early maintenance on the industrial equipment corresponding to the abnormal process subsequence to be detected, thereby realizing the prediction of early faults before the industrial equipment changes from a normal state to a faulty state, thereby taking preventive measures before the industrial equipment fault matures.

[0060] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for edge online processing of multi-source data, characterized in that: The following steps are involved: Step S1: According to the production plan information, the production process is divided into several process sub-sequences, edge monitoring points are set in each process sub-sequence, and data corresponding to each type of production data monitoring indicator of the process sub-sequence is obtained; Step S2: construct a relationship topology diagram between the industrial equipment corresponding to each process subsequence, set an edge data fuzzy judgment model at the edge monitoring point of each process subsequence, obtain the abnormal process subsequence according to the input edge data fuzzy judgment model, suspend the operation of the production equipment of the abnormal process subsequence, and start the edge scheduling mechanism of the abnormal process subsequence; Step S3: Build a full-process edge data prediction model based on the graph convolutional neural network, obtain the prediction data corresponding to each type of production data monitoring indicator of other edge monitoring nodes according to the full-process edge data prediction model, and determine whether each other edge monitoring node is an abnormal process subsequence to be detected.

2. The edge online processing method of multi-source data according to claim 1 is characterized in that: According to the production plan information, the production process is divided into several process sub-sequences, and edge monitoring points are set in each process sub-sequence. The process of obtaining data corresponding to each type of production data monitoring indicator of the process sub-sequence includes: Acquire the production plan information of the industrial production process of the current collection cycle, acquire the working time period and working order of each industrial equipment according to the production plan information, construct the production process sequence of the current collection cycle according to the working time period and working order of each industrial equipment, and segment the production process sequence according to the working time period of different industrial equipment in the production process sequence, and divide it into several process subsequences; Set edge monitoring points in each process subsequence to obtain the process flow characteristics of each industrial equipment, and use data retrieval to obtain various types of production data monitoring indicators of each edge monitoring point according to the process flow characteristics of each industrial equipment; The edge monitoring points acquire data corresponding to various types of production data monitoring indicators in real time according to various types of production data monitoring indicators, mark the collection time, and set the collection cycle.

3. The edge online processing method of multi-source data according to claim 2 is characterized in that: The process of constructing the relationship topology diagram between the industrial equipment corresponding to each process subsequence includes: Obtain the process flow characteristics of each industrial equipment, obtain the dependency relationship between the industrial equipment of each process subsequence according to the process flow characteristics of each industrial equipment, take each process subsequence as a node, take the dependency relationship between the industrial equipment of each process subsequence as the connection relationship between the nodes, and construct a relationship topology diagram; A local database is set at the edge monitoring point of each process subsequence, and the relationship topology map is stored in the local database of the data monitoring point of each process subsequence.

4. The edge online processing method of multi-source data according to claim 3 is characterized in that: An edge data fuzzy judgment model is set at the edge monitoring point of each process subsequence. The process of obtaining an abnormal process subsequence according to the input edge data fuzzy judgment model includes: An edge data fuzzy judgment model is set at the edge monitoring point of each process subsequence, and the data threshold interval corresponding to each type of production data monitoring indicator of each process subsequence is obtained in advance. The data threshold interval corresponding to each type of production data monitoring indicator of each process subsequence is taken as the fuzzy adaptive rule of the edge data fuzzy judgment model, and the data corresponding to each type of production data monitoring indicator collected in real time by the edge monitoring point of the process subsequence at the current moment is obtained. The data corresponding to each type of production data monitoring indicator is input into the edge data fuzzy judgment model of the edge monitoring point, and the edge data fuzzy judgment model compares the data corresponding to each type of production data monitoring indicator with the data threshold interval corresponding to each type of production data monitoring indicator; If the data corresponding to each type of production data monitoring indicator falls within the corresponding data threshold range, it proves that the production process of the industrial equipment corresponding to the process subsequence is normal; If the data corresponding to a certain type of production data monitoring indicator does not fall within the corresponding data threshold interval or the data corresponding to multiple types of production data monitoring indicators do not fall within the corresponding data threshold interval, the process subsequence is marked as an abnormal process subsequence.

5. The edge online processing method of multi-source data according to claim 4 is characterized in that: The process of suspending the operation of the production equipment of the abnormal process subsequence and starting the edge scheduling mechanism of the abnormal process subsequence includes: The monitoring center is communicated with the edge monitoring nodes of each process subsequence. When a process subsequence is marked as an abnormal process subsequence, the edge monitoring nodes of the abnormal process subsequence send abnormal monitoring indicator data to the monitoring center, and the monitoring center suspends the operation of the production equipment of the abnormal process subsequence. Subsequently, the edge monitoring points of the abnormal process subsequence obtain other process subsequences that have no connection relationship with the abnormal process subsequence according to the relationship topology diagram in the local database, obtain the remaining working period duration of the abnormal process subsequence after the suspension of operation and the working period duration of other process subsequences, screen out other process subsequences whose working period duration is less than or equal to the remaining working period duration of the abnormal process subsequence, randomly select one other process subsequence from the other process subsequences and mark it as the first process subsequence, obtain the working period and working order of the first process subsequence, and exchange the working period and working order of the abnormal process subsequence with the working period and working order of the first process subsequence.

6. The edge online processing method of multi-source data according to claim 5, characterized in that: The process of building a full-process edge data prediction model based on the graph convolutional neural network and obtaining the prediction data corresponding to various types of production data monitoring indicators of other edge monitoring nodes according to the full-process edge data prediction model includes: The monitoring center stores historical data corresponding to various types of production data monitoring indicators stored in several historical collection cycles of each edge monitoring node, and uses the historical data corresponding to various types of production data monitoring indicators stored in several historical collection cycles of each edge monitoring node as a supplementary node to the node to which each edge monitoring node belongs in the relationship topology diagram; According to the adjacency matrix of each node in the relationship topology graph, the influence degree of dependency on other nodes is calculated, and the weight label is set for the node by obtaining the influence degree of dependency of each node on other nodes; Subsequently, a full-process edge data prediction model is constructed based on the graph convolutional neural network, the relationship topology graph and the supplementary nodes in the relationship topology graph are learned, and the data corresponding to each type of production data monitoring indicators of the abnormal process subsequence are input into the learned full-process edge data prediction model. According to the full-process edge data prediction model, the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes are obtained, and the monitoring center sends the prediction data corresponding to each type of production data monitoring indicators of other edge monitoring nodes to the corresponding edge monitoring nodes.

7. The edge online processing method of multi-source data according to claim 6 is characterized in that: According to the prediction data corresponding to the various types of production data monitoring indicators of the other edge monitoring nodes, the process of judging whether the other edge monitoring nodes are abnormal process subsequences to be detected includes: Each other edge monitoring node inputs the received prediction data corresponding to each type of production data monitoring indicator into the edge data fuzzy judgment model contained in each other edge monitoring node, and the edge data fuzzy judgment model compares the prediction data corresponding to each type of production data monitoring indicator with the data threshold interval corresponding to each type of production data monitoring indicator; If the predicted data corresponding to a certain type of production data monitoring indicator does not fall within the corresponding data threshold interval or the predicted data corresponding to multiple types of production data monitoring indicators do not fall within the corresponding data threshold interval, the process subsequence is marked as an abnormal process subsequence to be detected.

Images